3. Stacked densities¶
3.1. 概要¶
3.2. Plotlyによる作図方法¶
3.3. MADB Labを用いた作図例¶
3.3.1. 下準備¶
import pandas as pd
import plotly.express as px
import warnings
warnings.filterwarnings('ignore')
# 前処理の結果，以下に分析対象ファイルが格納されていることを想定
PATH_DATA = '../../data/preprocess/out/magazines.csv'
# Jupyter Book用のPlotlyのrenderer
RENDERER = 'plotly_mimetype+notebook'
def show_fig(fig):
"""Jupyter Bookでも表示可能なようRendererを指定"""
fig.update_layout(margin=dict(t=50, l=25, r=25, b=25))
fig.show(renderer=RENDERER)
def add_years_to_df(df, unit_years=10):
"""unit_years単位で区切ったyears列を追加"""
df_new = df.copy()
df_new['years'] = \
pd.to_datetime(df['datePublished']).dt.year \
// unit_years * unit_years
df_new['years'] = df_new['years'].astype(str)
return df_new
df = pd.read_csv(PATH_DATA)
3.3.2. 雑誌別・年代別の合計作品数¶
col_count = 'cname'
# 1年単位で区切ったyearsを追加
df = add_years_to_df(df, 1)
# mcname, yearsで集計
df_plot = \
df.groupby(['mcname', 'years'])[col_count].\
nunique().reset_index()
# years単位で集計してdf_plotにカラムを追加
df_tmp = df_plot.groupby('years')[col_count].sum().reset_index(
name='years_total')
df_plot = pd.merge(df_plot, df_tmp, how='left', on='years')
# years合計あたりの比率を計算
df_plot['ratio'] = df_plot[col_count] / df_plot['years_total']
fig = px.area(
df_plot, x='years', y='ratio', color='mcname',
title='雑誌別・年代別の合計作品数')
show_fig(fig)
3.3.3. 雑誌別・年代別の合計作者数¶
col_count = 'creator'
# 10年単位で区切ったyearsを追加
df = add_years_to_df(df, 1)
# mcname, yearsで集計
df_plot = \
df.groupby(['mcname', 'years'])[col_count].\
nunique().reset_index()
# years単位で集計してdf_plotにカラムを追加
df_tmp = df_plot.groupby('years')[col_count].sum().reset_index(
name='years_total')
df_plot = pd.merge(df_plot, df_tmp, how='left', on='years')
# years合計あたりの比率を計算
df_plot['ratio'] = df_plot[col_count] / df_plot['years_total']
fig = px.area(
df_plot, x='years', y='ratio', color='mcname',
title='雑誌別・年代別の合計作者数')
show_fig(fig)